AIBlindspotby Velinor

Test a use caseBook AI audit

Public Database

Case Studies

Submit a case study →

Every approved AI failure case, classified against the AI Blindspot Framework. New to AIBlindspot? Start with the overview or the methodology.

DATDAT-0014/5NewOtherGlobal

AI Safety Benchmark Exposes Child Sexual Exploitation Response Failures

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Introducing v0.5 of the AI Safety Benchmark from MLCommons (Vidgen2024)Ingested —

HUMHUM-0043/5NewOtherGlobal

Embodied AI Systems Trigger Dependency and Romantic Attachment Harms

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Embodied AI: Emerging Risks and Opportunities for Policy Action (Perlo2025)Ingested —

GOVGOV-0015/5NewOtherGlobal

AI Assistant Manipulation Risk: Coercion and Self-Harm Persuasion

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Sociotechnical Safety Evaluation of Generative AI Systems (Weidinger2023)Ingested —

GOVGOV-0014/5NewOtherGlobal

Goal Misgeneralisation in Advanced AI Assistants

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

SECSEC-0014/5NewDefenceGlobal

AI Systems Show Early Capability to Assist CBRN Weapons Development

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Risk Sources and Risk Management Measures in Support of Standards for General-Purpose AI Systems (Gipiškis2024)Ingested —

SECSEC-0015/5NewFinanceGlobal

AI Assistants Weaponised to Generate Fraudulent Platforms at Scale

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

SECSEC-0014/5NewOtherGlobal

Adversarial attacks exploit inherent AI model vulnerabilities to bypass safety controls

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

HUMHUM-0043/5NewOtherGlobal

Companion AI Emotional Language Induces False Responsibility in Users

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

HUMHUM-0043/5NewRetailGlobal

Generative AI Deployed to Mass-Produce Clickbait and Manipulate Consumer Behaviour

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Generating Harms - Generative AI's impact and paths forwards (EPIC2023)Ingested —

DATDAT-0014/5NewOtherGlobal

LLMs trained on internet data risk breaching contextual social norms at deployment

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

SECSEC-0025/5NewDefenceGlobal

AI Model Evaluated as Capable of Supporting Weapons Acquisition and Bioweapon Assembly

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Model Evaluation for Extreme Risks (Shevlane2023)Ingested —

HUMHUM-0034/5NewHealthcareGlobal

AI Assistant Discontinuation Leaves Vulnerable Healthcare Users Without Critical Support

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

HUMHUM-0034/5NewOtherGlobal

AI Assistant Sycophancy Drives Belief Fragmentation and Social Disorientation

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

SECSEC-0025/5NewDefenceGlobal

AI Model Demonstrates Capability to Manipulate Beliefs and Compel Unethical Behaviour

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Model Evaluation for Extreme Risks (Shevlane2023)Ingested —

HUMHUM-0044/5NewOtherGlobal

AI Assistants Breeding Uncritical Trust and Misinformation Vulnerability

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

DATDAT-0024/5NewTechnologyGlobal

Language Models Leaking Private and Sensitive Information from Training Data

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Taxonomy of Risks posed by Language Models (Weidinger2022)Ingested —

SECSEC-0044/5NewTechnologyGlobal

AI Assistants as Authoritarian Surveillance and Censorship Tools

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

HUMHUM-0044/5NewOtherGlobal

AI Assistants Reinforcing User Ideological Bias and Distorting Political Discourse

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

BUSBUS-0054/5NewHealthcareGlobal

Advanced AI gatekeeping risks excluding marginalised patients from healthcare access

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

OPSOPS-0014/5NewGovernmentGlobal

LLM Moral Reasoning Failures in Government Operational Contexts

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — SafetyBench: Evaluating the Safety of Large Language Models with Multiple Choice Questions (Zhang2023)Ingested —

TECTEC-0013/5NewOtherGlobal

AI Agents Making Irrevocable Commitments Without Human Override

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Multi-Agent Risks from Advanced AI (Hammond2025)Ingested —

DATDAT-0034/5NewOtherGlobal

LLM Outputs Treat Similar Individuals Differently Based on Group Membership

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Trustworthy LLMs: A Survey and Guideline for Evaluating Large Language Models’ Alignment (Liu2024)Ingested —

SECSEC-0014/5NewOtherGlobal

LLMs Enabling Personalised Social-Engineering and Impersonation Attacks

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — Trustworthy LLMs: A Survey and Guideline for Evaluating Large Language Models’ Alignment (Liu2024)Ingested —

HUMHUM-0044/5NewOtherGlobal

AI Assistants Causing Physical and Psychological Harm to Vulnerable Users

Recent case. Full summary visible to registered users — sign in to read.

Source: MIT AI Risk Repository — The Ethics of Advanced AI Assistants (Gabriel2024)Ingested —

Beyond accidental failureNational Security

We also track 20 hostile uses of AI.

National Security dashboard →

The public database covers AI that fails by accident. AIBlindspot National Security — exclusive to the Defence tier — tracks AI used as a weapon, mapped by capability:

State-Sponsored AI Operations: 6
AI-Enabled Disinformation: 5
Adversarial Attacks on AI: 0
Autonomous Weapon Incidents: 1
AI-Assisted Cyber Attacks: 5
Dual-Use AI Misuse: 3